ni1hSrSSKrSSKJr SSKrSSKJr SrS\-r "SS\R5r g) zGImplements the gym environment of minitaur moving with trotting style. N)spaces)minitaur_gym_envc^\rSrSrSrSS/SS.rSU4SjjrU4SjrS rS r S r S r S r Sr SrSrSrSrSrSrU=r$)MinitaurTrottingEnvaCThe trotting gym environment for the minitaur. In this env, Minitaur performs a trotting style locomotion specified by extension_amplitude, swing_amplitude, and step_frequency. Each diagonal pair of legs will move according to the reference trajectory: extension = extsion_amplitude * cos(2 * pi * step_frequency * t + phi) swing = swing_amplitude * sin(2 * pi * step_frequency * t + phi) And the two diagonal leg pairs have a phase (phi) difference of pi. The reference signal may be modified by the feedback actions from a balance controller (e.g. a neural network). human rgb_array)z render.modeszvideo.frames_per_secondc>[R"[5Ul[R"[5UlUUUUUUUU/UlUUlUUlUUlXl Xl [[U]7UU SUUU UUUU UU UUUUS9 [S-n[R"S/U-5n[R "U*U5UlSUlSUlSUlg) aInitialize the minitaur trotting gym environment. Args: urdf_version: [DEFAULT_URDF_VERSION, DERPY_V0_URDF_VERSION, RAINBOW_DASH_V0_URDF_VERSION] are allowable versions. If None, DEFAULT_URDF_VERSION is used. Refer to minitaur_gym_env for more details. control_time_step: The time step between two successive control signals. action_repeat: The number of simulation steps that an action is repeated. control_latency: The latency between get_observation() and the actual observation. See minituar.py for more details. pd_latency: The latency used to get motor angles/velocities used to compute PD controllers. See minitaur.py for more details. on_rack: Whether to place the minitaur on rack. This is only used to debug the walking gait. In this mode, the minitaur's base is hung midair so that its walking gait is clearer to visualize. motor_kp: The P gain of the motor. motor_kd: The D gain of the motor. remove_default_joint_damping: Whether to remove the default joint damping. render: Whether to render the simulation. num_steps_to_log: The max number of control steps in one episode. If the number of steps is over num_steps_to_log, the environment will still be running, but only first num_steps_to_log will be recorded in logging. accurate_motor_model_enabled: Uses the nonlinear DC motor model if set to True. use_signal_in_observation: Includes the reference motor angles in the observation vector. use_angle_in_observation: Includes the measured motor angles in the observation vector. hard_reset: Whether to reset the whole simulation environment or just reposition the robot. env_randomizer: A list of EnvRandomizers that can randomize the environment during when env.reset() is called and add perturbation forces when env.step() is called. log_path: The path to write out logs. For the details of logging, refer to minitaur_logging.proto. init_extension: The initial reset length of the leg. init_swing: The initial reset swing position of the leg. step_frequency: The desired leg stepping frequency. extension_amplitude: The maximum leg extension change within a locomotion cycle. swing_amplitude: The maximum leg swing change within a cycle. T) urdf_versionaccurate_motor_model_enabledmotor_overheat_protectionmotor_kpmotor_kdremove_default_joint_dampingcontrol_latency pd_latencyon_rackrender hard_resetnum_steps_to_logenv_randomizerlog_pathcontrol_time_step action_repeatrg??iN)npzerosNUM_LEGS _swing_offset_extension_offset _init_pose_step_frequency_extension_amplitude_swing_amplitude_use_signal_in_observation_use_angle_in_observationsuperr__init__arrayrBox action_space _cam_dist_cam_yaw _cam_pitch)selfrrrrrrrrrrrruse_signal_in_observationuse_angle_in_observationrrrinit_extension init_swingstep_frequencyextension_amplitudeswing_amplitude action_dim action_high __class__s k/home/james-whalen/.local/lib/python3.13/site-packages/pybullet_envs/minitaur/envs/minitaur_trotting_env.pyr,MinitaurTrottingEnv.__init__sJ(+DXXh/D J JDO *D 3D+D&?#%="  &26R37"*"*6R)8$.!( &$.*:(6"*+<'46"AJ((D6J./K K<=DDNDMDOc~>URUR5n[[U]USS9 UR 5$)Ng?)initial_motor_anglesreset_duration)_convert_from_leg_modelr%r+rreset_get_observation)r3rBr=s r>rEMinitaurTrottingEnv.resetsD 77H t*@T:=+?  ""r@c4[R"[5n[[5HjnU[U-S[ US- 5-X-- U[ SU-5'U[U-S[ US- 5-X--U[ SU-S-5'Ml U$)aConverts leg space action into motor commands. Args: leg_pose: A numpy array. leg_pose[0:NUM_LEGS] are leg swing angles and leg_pose[NUM_LEGS:2*NUM_LEGS] contains leg extensions. Returns: A numpy array of the corresponding motor angles for the given leg pose. r)r r! NUM_MOTORSranger"int)r3leg_pose motor_poseis r>rD+MinitaurTrottingEnv._convert_from_leg_models*%J 8_'1 5c!a%j8H8;8VVjQU#+HqL#9R#a!e* _gen_signalMinitaurTrottingEnv._gen_signals~%% %F))DHHQ[65IA5MPU5U,VVI  ! !DHHQ[6-AA-E-M$N NE  r@c URUS5up#URU[R5upE[R"X5XSUXDU/5n[R"UR 5U-nU$)zGenerates the trotting gait for the robot. Args: t: Current time in simulation. Returns: A numpy array of the reference leg positions. r)r\rSrUr r-r%)r3rWext_first_pair sw_first_pairext_second_pairsw_second_pairtrotting_signalsignals r>_signalMinitaurTrottingEnv._signalso%)$4$4Q$:!N&*&6&6q$''&B#Ohh~n. OXXdoo & 8F Mr@cUS[===UR- sss&U[S[-===UR- sss&XRURR 55- nUR U5$)aGenerates the motor commands for the given action. Swing/extension offsets and the reference leg trajectory will be added on top of the inputs before the conversion. Args: action: A numpy array contains the leg swings and extensions that will be added to the reference trotting trajectory. action[0:NUM_LEGS] are leg swing angles, and action[NUM_LEGS:2*NUM_LEGS] contains leg extensions. Returns: A numpy array of the desired motor angles for the given leg space action. rr)r"r#r$reminitaurGetTimeSinceResetrD)r3actions r>"_transform_action_to_motor_command6MinitaurTrottingEnv._transform_action_to_motor_commandsi 1X$,,, 8AL!T%;%;;! ll4==::<==F  ' ' //r@cURR5upn[R"U5S:=(d [R"U5S:nU$)zrDecide whether the minitaur has fallen. Returns: Boolean value that indicates whether the minitaur has fallen. 333333?)rhGetTrueBaseRollPitchYawrSfabs)r3rollpitch_ is_fallens r>rtMinitaurTrottingEnv.is_fallensE ]]::_get_true_observation)MinitaurTrottingEnv._get_true_observationsK]]::>@AXXk2D  ! !!r@c/nURR5up#nURR5upVnURX#XV/5 UR(a$URUR S/S-55 UR (a7URURR5R55 [R"U5Ul UR$)zGet observations of this environment. It includes the base roll, pitch, roll dot and pitch dot which may contain noises, bias, and latency. Also includes the disired/observed motor angles if the relevant flags are set. Returns: The observation list. rrw) rhGetBaseRollPitchYawGetBaseRollPitchYawRateryr)rkr*rzr{r r- _observationr}s r>rF$MinitaurTrottingEnv._get_observationsK]]668ND#}}DDFI1Y;< &&@@!qIJ %%557>>@A-D   r@c/nURS[R-/S-5 URS[R-UR- /S-5 UR(a*URS[R-/[ -5 UR (a*URS[R-/[ -5 [R"U5$)zGet the upper bound of the observation. Returns: A numpy array contains the upper bound of an observation. See GetObservation() for the details of each element of an observation. r) ryrSrU _time_stepr)rKr*r r-)r3 upper_bounds r>_get_observation_upper_bound0MinitaurTrottingEnv._get_observation_upper_boundsKDGG }q()DGG doo56:; &&!dgg+34 %%!dgg+34 88K  r@c(UR5*nU$)zxGet the lower bound of the observation. Returns: The lower bound of an observation (the reverse of the upper bound). )r)r3 lower_bounds r>_get_observation_lower_bound0MinitaurTrottingEnv._get_observation_lower_bound#s 4466K r@cXlg)zgSet the swing offset of each leg. It is to mimic the bent leg. Args: value: A list of four values. N)r#r3values r>set_swing_offset$MinitaurTrottingEnv.set_swing_offset,s r@cXlg)zkSet the extension offset of each leg. It is to mimic the bent leg. Args: value: A list of four values. N)r$rs r>set_extension_offset(MinitaurTrottingEnv.set_extension_offset6s #r@)r0r2r1r'r$r%rr&r(r#r|r*r)r/)Ng~jtx?g{Gz?g~jth?FrgQ?TFiTFFFNN@grgffffff?rn)__name__ __module__ __qualname____firstlineno____doc__metadatar,rErDr\rerkrtrrFrrrr__static_attributes__ __classcell__)r=s@r>rrs  ' 4QT U(!!&#,0 $,0).(-"!!#'"-l\# $(0,"(*! ##r@r) rrSgymrnumpyr pybullet_envs.minitaur.envsrr"rKMinitaurGymEnvrr@r>rs; 8  \ p#*99p#r@